Molecular Biology of the Cell Vol. 3, 1279-1294, November 1992

Identification of the Domains in Cyclin A Required for Binding to, and Activation of, p34cdC2 and p32cdk2 Protein Kinase Subunits Hideki Kobayashi,* Elspeth Stewart, Randy Poon, Jorg P. Adamczewski, Julian Gannon, and Tim Huntt ICRF Clare Hall Laboratories, South Mimms, Herts EN6 3LD, England Submitted July 7, 1992; Accepted September 14, 1992

The binding of cyclin A to p34cdc2 and p32cdk2 and the protein kinase activity of the complexes has been measured by cell-free translation of the corresponding mRNA in extracts of frog eggs, followed by immunoprecipitation. A variety of mutant cyclin A molecules have been constructed and tested in this assay. Small deletions and point mutations of highly conserved residues in the 100-residue "cyclin box" abolish binding and activation of both p34cdc2 and p32cdk2. By contrast, large deletions at the N-terminus have no effect on kinase binding and activation, until they remove residues beyond 161, where the first conserved amino acids are found in all known examples of cyclin A. At the C-terminus, removal of 14 or more amino acids abolishes activity. We also demonstrate that deletion of, or point mutations, in the cyclin A homologue of the 10-residue "destruction box," previously described in cyclin B (Glotzer et al., 1991) abolish cyclin proteolysis at the transition from M-phase to interphase. INTRODUCTION

Cell cycle transitions require the activity of protein kinases that contain catalytic subunits encoded by members of the cdc2 gene family together with a regulatory and activating subunit corresponding to a member of the cyclin family (reviewed by Nurse (1990) and Pines and Hunter (1990). The association of a mitotic cyclin with p34cdc2 appears to be necessary to turn on its protein kinase activity (Solomon et al., 1990; Desai et al., 1992), and the destruction of cyclin turns off the protein kinase activity of p34cdc2 just before the onset of anaphase (Luca and Ruderman, 1989; Murray et al., 1989; Luca et al., 1991). It is now known that there is more than one cdc2 gene in higher eukaryotes (Lehner and O'Farrell, 1990; Elledge and Spottswood, 1991; Paris et al., 1991; Tsai et al., 1991) and at least eight different kinds of cyclin (mitotic cyclins A, Bi, and B2 in vertebrates; G1 cyclins CLN1/2 and CLN3 in budding yeast, and three cyclin homologues known as cyclins C, D (CYLl or PRAD1) and E have recently been discovered * Permanent address: Department of Biology, Faculty of Science, Kyushu University, Fukuoka 812, Japan. t To whom correspondence should be addressed.

© 1992 by The American Society for Cell Biology

in humans (Hadwiger et al., 1989; Leopold and O'Farrell, 1991; Lew et al., 1991; Matsushime et al., 1991; Motokura et al., 1991; Xiong et al., 1991). It is not yet clear what combinations of cyclin plus p34cdc2-like subunits can associate to form active protein kinases or what special properties different cyclin subunits confer on their partners. Indeed, for some of the most recently discovered cyclins, particularly C and D, the evidence for association with protein kinase subunits is at present somewhat indirect. It is thought likely that cyclins are in some sense targeting subunits for p34cdc2 as well as controlling the timing of turning on and off its kinase activity (Minshull et al., 1990; Solomon et al., 1990). So far, most of the known substrates for cyclin/cdc2 protein kinases are able to be phosphorylated by any combination of cyclin and cdc2 subunit (Minshull et al., 1990) provided their kinase activity has been turned on by the appropriate phosphorylation state of the cdc2 subunit (Dunphy and Newport, 1989; Gould and Nurse, 1989; Morla et al., 1989; Newport, 1989; Ducommun et al., 1991; Featherstone and Russell, 1991; Gautier et al., 1991; Gould et al., 1991; Parker et al., 1991, 1992; Strausfeld et al., 1991; Kumagai and Dunphy, 1992; Millar and Russell, 1992) (for reviews see Fleig and Gould, 1991 and Maller, 1991). The cellular targets of 1279

H. Kobayashi et al.

these kinases are not well characterized, however, and at the relatively low concentrations of kinases and their target substrates found in cells, the specificity may be much higher. The sequence homology between different members of the cyclin family is not extensive and is largely confined to a stretch of 100 residues in the middle of the linear sequence that is commonly known as the "cyclin box" (Nugent et al., 1991). Even in this core, however, only five well-spaced residues (RDLKF) show complete conservation in all cyclins from yeast to man, and in cyclin C, which appears to be a yet more distant relative, the R and D are altered. If, as is believed, these proteins all associate with p34cdc2 or p34cdc2 -like proteins, it is likely that these conserved residues and the region spanned by them are important in the cyclin-p34cdc2 interaction, and if this is so, questions are raised about the function of the remaining 300 or so amino acids. One other clearly marked domain has been defined in the N-terminus of the mitotic cyclins A and B, which is required for their regulated sudden destruction at the metaphase-anaphase transition (Murray et al., 1989; Glotzer et al., 1991; Lorca et al., 1992). As a first step to investigating structure-function relationships in cyclin A, we have made a number of N- and C-terminal deletions of Xenopus and bovine cyclin Al and various small deletions and point mutations in the cyclin box. We have tested these constructs for their ability to promote Xenopus oocyte maturation, for their ability to bind and activate p34cdc2 and p32cdk2, and for their ability to be destroyed. We find that the first 161 residues of cyclin A are dispensable for binding to and activation of p34cdc2, but larger deletions, which encroach on the first conserved residues, cause complete loss of activity. By contrast, removal of as little as 14 amino acids at the C-terminus of cyclin A renders it completely inactive. Point mutations of the completely conserved R197 and D226 and deletions of two or more residues from the cyclin box cause serious or complete. loss of activity. There is a perfect correlation between the ability of cyclin A to bind with p34cdc2 and its activity, both in vivo and in vitro. We also confirm that the N-terminal destruction box in cyclin A is necessary for programmed destruction triggered by addition of Ca21 to an egg extract (Glotzer et al., 1991; Luca et al., 1991; Lorca et al., 1992). Finally, although we and others were previously unable to detect newly synthesized cdclin A associated with immunoprecipitates of p32 cdk (Minshull et al., 1990; Gabrielli et al., 1992), we now report that when Xenopus cyclin Al and p32cdk2 are capable of binding together to give an active histone kinase when their concentrations are increased by cell-free translation of added mRNA. By contrast, B-type cyclins do not appear to form complexes with p32cd 2 under the same conditions. The cyclin A mutations that abolish p34cdc2 binding also abolish p32cdk2 binding. 1280

MATERIALS AND METHODS Construction of Cyclin A Deletions A full-length clone of Xenopus cyclin Al1 in pGEM1 (clone XL4) (Minshull et al., 1990) was digested at the unique EcoRV site. The linearized DNA (10 ,Ag) was incubated with 5 U of Bal 31 nuclease (Boehringer, Mannheim, Germany) in a final volume of 100 ,sl in 20 mM tris(hydroxymethyl)aminomethane (Tris)2-Cl, 100 mM NaCl, 5 mM CaCl2, 5 mM MgCl2, 1 mM EDTA, pH 8.0, at 30°. Samples of 15 Ml were taken at 2-min intervals into an equal volume of 20 mM ethylene glycol-bis(,-aminoethyl ether)-N,N,N',N'-tetraacetic acid) (EGTA), pH 7.3, to stop further digestion. The DNA was repaired with T4 DNA polymerase in the presence of 100 ,uM dNTPs, ligated with T4 DNA ligase, and introduced into Escherichia coli strain TG1 by standard Ca2"-mediated transformation. DNA was prepared from individual colonies and the size of deletion assessed by restriction enzyme analysis followed by agarose gel electrophoresis. Selected clones were transcribed (see below) and translated in the nuclease-treated rabbit reticulocyte lysate, and the [35S]methionine-labeled products analyzed

by sodium dodecylsulfate-polyacrylamide gel electrophoresis (SDSPAGE). The clones that gave translation products of the expected size (as opposed to prematurely terminated products resulting from outof-frame deletions) were selected for further analysis and were sequenced with an oligonucleotide with the sequence 5' ACAAGCTGAAGTTTTCC 3', corresponding to the amino acid sequence GKLQLV in Xenopus cyclin Al. Similar deletions were made from the Nsi I and Sty I sites (we thank Hashmat Sikder, who isolated and characterized A62). The junction sequences of the selected clones are listed in Table 1.

C-Terminal Cyclin Deletions A series of C-terminal cyclin deletions, all of them including a terminator codon, were constructed by the polymerase chain reaction, with 3' primers containing a BamHI site, which was used to insert the truncated DNA between the HindIIl and Bcl I sites of a modified cyclin Al construct in which the first 21 residues of cyclin Al were replaced with 15 amino acids of the c-myc epitope described by Evan et al. (1986), and the translational leader was derived from influenza virus NS protein as described by Dasso and Jackson (1989). The plasmid backbone was derived from pGEM2 (Promega, Madison, WI), modified by removal of the 252 bp EcoRI-Nae I fragment, in whose place was added an EcoRI-EcoRV fragment from a pET8 construct that included the T7 transcription terminator (the idea was to avoid the linearization step in the transcription protocol, but we found that there was too much readthrough, so we did not exploit this feature in practice). The sequence from the start of the T7 promoter to the start of the cyclin sequence was as follows TAATACGACTCACTATAGGGAGACCGGAAGCTAGCTTGGGCTGCAGGTCGACAGCAAAAGCAGGGTGAC CAAAGACATAATGGATCCCATGGAGCAAAAGCTCATTTCTGAAGAAGATCTGAACAGTGCATTCCAGAA M D P M E O K L

I

S E E D L N S A F Q N

The c-myc tag sequence is underlined. In this construct there is a unique EcoRI site at the 3' end of the cyclin 3' UTR (untranslated region), and the Nhe I, BstEII, BamHI, Nco I, and Bgl II in the sequence above are all unique in the plasmid (addition of the T7 terminator

1 We refer to cyclin Al to describe the form of Xenopus cyclin A used throughout this paper. Xenopus has at least one other cyclin A gene, which may have different properties (M. Howell and T.H., unpublished data). 2 Abbreviations used: CSF, cytostatic factor; EGTA, ethylene glycolbis(f-aminoethyl ether)-N,N,N',N'-tetraacetic acid; GVBD, germinal vesicle breakdown; HEPES, N-2-hydroxyethylpiperazine-N'-2-ethanesulfonic acid; IgG, immunoglobulin G; PAGE, polyacrylamide gel electrophoresis; SDS, sodium dodecyl sulfate; Tris, tris(hydroxymethyl)aminomethane; UTR, untranslated region of mRNA. Molecular Biology of the Cell

Functional Domains of Cyclin A

Table 1. Details of cyclin deletion constructs Name

Length of deletion

Location of deletion

Mutants of Xenopus cyclin Al A21 21 1-21 A62 39 24-62 A133 133 1-133 A144 57 88-144 A147 58 90-147 A158 57 102-158 A161 53 109-161 A169 69 101-169 A176 84 93-176 A201 122 80-201 A213 117 97-213 A245 210 36-245 A232 2 231-232 A235 7 229-235 A237 10 228-237 A261 36 226-261 CA14 14 405-419 CA24 24 395-419 CA50 50 369-419 CA79 79 340-419 CA200 200 219-419

Junction sequence

MEQKLISEEDLNaSAF ASSA/PAKS MDPMbEASP NPAP/QTSP APVA/PEDD VDEP/VAVS YSVE/SEYI YVDE/QYLR APKP/LKHR AFPG/MCDWL SFTV/HTET VEVQ/AILL FSLS/SVLR LDRF/RGKL YLDR/KLQL MNYL/DEFV KTTKYM*d QAQQAI* AFTGYA* KYVPSL* LHTETL*

Mutants of Bovine cyclin Ae A267 A275 A278 CA16

2 18 24 16

266-267 258-275 255-278 403-419

AEFV/SfDDT EEIY/QVLR SKFE/GgMEH REKYKNS*

a Underlined sequence denote the c-myc tag. This leader was present in all the Xenopus C-terminal deletion series of constructs except

CA200.

b Italics indicate altered residues from vector sequences; the wild-type

sequence is LDIS. M replaces V. d An asterisk denotes a termination codon. eThe numbering system is that of the equivalent Xenopus sequences. 'The S is a T in the wild-type sequence. 8 The G is an R in the wild-type seuqence.

removed an Nhe I site in the vector), as are the internal EcoRV, HindIII, Nsi I, and Bcl I sites present in Xenopus cyclin Al. The C-terminal deletion CA200 was derived from the Nsi I series of Bal 31 deletions. Closure of the deleted molecule formed a termination codon (Table 1).

Construction of Point Mutations Mutagenesis of individual residues in cyclin Al was either performed with the Amersham (Arlington Heights, IL) kit, or with a polymerase chain reaction-based strategy essentially as described by Horton and Pease (1991). The altered segments were checked by sequencing the double-stranded template DNA with the United States Biochemicals Sequenase kit (Cleveland, OH).

a bovine lymphocyte cDNA library.3 The large Sau3AI fragment from this clone was inserted into the in-frame BamHI site of a vector derived from pET3a (Studier et al., 1990) encoding the IgG (immunoglobulin G) binding domain of staphylococcal protein A (derived by PCR from pRIT2T [Pharmacia, Uppsala, Sweden]). This construct encoded protein A at the N-terminus and bovine cyclin A at the C-terminus, minus the first 76 residues of the cyclin A (the junction sequence is FIQSLKDDPGNSRDLPINDEYVPVPP, cyclin residues underlined). The fusion protein comprised 620 amino acids and was soluble when induced with 0.1 mM isopropyl-A-D galactoside (IPTG) at an A600 of 0.4 in E. coli BL21 (DE3) at 200 overnight. After a freeze-thaw cycle, the bacteria were digested with lysozyme, and the protein A-cyclin A constructs were purified from the 100 000 g supematant by chromatography on IgG Sepharose as described by Solomon et al. (1990). Mutations were introduced into the construct by Bal 31 deletions at a unique Acc I site, and deletion of 16 residues at the C-terminus was performed by means of the polymerase chain reaction. The structures of these constructs are summarized in Table 1 and Figure 5.

Preparation of mRNA for Translation Plasmid DNA was linearized at the end of the 3' UTR with the ap-

propriate restriction enzyme (BamHI for the untagged cyclin A constructs, EcoRI for the tagged ones) and used as templates for transcription in vitro using T7 RNA polymerase essentially according to Nielsen and Shapiro (1986). Capped mRNAs were made in 50-pl reactions that contained 4-5 ug of DNA, 40 mM Tris-Cl, pH 8.0, 15 mM MgCl2, 5 mM dithiothreitol (DTT), 1 mM each of ATP, cytidine

triphosphate (CTP), and uridine triphosphate (UTP), 0.1 mM guanosine triphosphate (GTP), 0.5 mM m7G(5')ppp(5')G (New England Biolabs, Beverly, MA), 50 units of RNasin (Boehringer), and 25 units of T7 RNA polymerase (purified from E. coli BL21 harboring pAR1219 according to a protocol supplied by Dr. J.J. Dunn, Brookhaven National Laboratory, Upton, New York; [see Studier et al., 1990] or purchased from New England Biolabs). After 30 min at 370, the GTP concentration was increased to 1 mM and incubation continued for a further 60 min. The reactions were extracted twice with phenol/chloroform, and the RNA was recovered by ethanol precipitation. The RNA pellets were resuspended in 0.2 mM EDTA, pH 7.0, at a concentration of 12 ,ug/,ul.

Microinjection of mRNA into Xenopus Oocytes Stage VI oocytes were obtained from female Xenopus laevis, with the use of mild digestion with collagenase as described by Colman (1984). We injected -50 nl of 1 ng/nl mRNA into each oocyte, which were kept at 220 in modified Barth's medium. Maturation was assessed by

the appearance of the characteristic white spot on the animal pole and checked in doubtful cases by dissection.

Translation in Cell-Free Extracts of Xenopus Eggs For cell-free translation of mRNA, we prepared "cytostatic factor (CSF)-arrested" extracts from unactivated Xenopus eggs according to the protocol of Murray (1991). Sucrose was added to the extracts to a final concentration of 200 mM, and small aliquots were frozen in liquid nitrogen. For translation, mRNA was added at 100 ,ug/ml and [35S]methionine to 0.5 mCi/ml in a reaction of 10 ,ul, which was incubated at 230 for 1 h. Part of the reaction mix was used for SDSPAGE and autoradiography (typically, the equivalent of 0.2-0.5 Al was analyzed on one gel lane) and another part for immunoprecipitation or p133ucl affinity chromatography as described below. In some cases, the frog cell extract was mixed with an equal volume of nucleasetreated rabbit reticulocyte lysate, which stimulated translation of certain added mRNAs, some of which (for example the mRNAs for p34cdc2

Construction and Expression of Bovine Cyclin A A partial clone encoding all but the first 24 residues (compared with the sequence of human cyclin A) of bovine cyclin A was isolated from

Vol. 3, November 1992

3 The sequence of this bovine cyclin A clone has been deposited in the GENBANK/EMBL database with accession number X68321.

1281

H. Kobayashi et al.

and p32cdk2) were translated extremely poorly in CSF extracts. We found that placing the coding regions of such mRNAs within the 5' and 3' untranslated regions of cyclin Al mRNA improved their translatability. Conversely, removing the 3' UTR from cyclin Al mRNA strongly reduced its translation in the Xenopus cell-free system, although it did not significantly reduce its translation in the reticulocyte system. We estimated the concentration of proteins produced by cell-free translation of added mRNA in pure reticulocyte lysate in two independent ways: by measuring the radioactivity in the protein, knowing the specific activity of the methionine pool, and also by immunoblotting and comparison with bacterial cyclin A standards. The two methods gave estimates of 0.4-1 Mg/ml for wild-type cyclin A after 60 min translation. We did not perform these measurements in frog extracts.

Affinity Chromatography on p13s"cl-Sepharose The affinity matrix for p34 '2, pl3sucI-Sepharose, was prepared by modifications of the procedure described by Brizuela et al. (1987). The pl3sucl was purified from E. coli BL21(DE3) carrying the intronless p13Sucl gene in the modified pET3 vector pRK172, grown in 2xTY medium and induced with 0.5 mM IPTG. The bacteria were lysed on ice in 20 mM HEPES (N-2-hydroxyethylpiperazine-N'-2-ethanesulfonic acid), 2 mM EDTA, 1 mM DTT, and 1 mM PMSF, pH 7.2, by addition of 2 mg/ml lysozyme followed by sonication. The extract was clarified by centrifugation for 20 min at 10 000 rpm in the Sorvall SS34 rotor (Sorvall Instruments, Newton, CT), dialyzed against the lysis buffer, and loaded on a DEAE-Sepharose column. The column was developed with a 0-500 mM NaCl gradient; p13sucl eluted between 30 and 60 mM NaCl, well ahead of most of the other proteins. The pooled fractions containing the p13SucI were concentrated with 65% saturated ammonium sulfate, and the pellet was dissolved in coupling buffer: 500 mM NaCl, 100 mM Na2CO3. This solution was then further purified by gel filtration on AcA 34 (IBF Biotechnics, Villeneuve-la-Garenne, France) equilibrated with the NaCl/carbonate coupling buffer. This p13SUcl appeared extremely pure by SDS-PAGE, and the peak column fractions were used directly for coupling to the

beads.

For analysis of translation reactions, 5 Mil were diluted with 400 ,l of 'bead buffer' (Dunphy et al., 1988): 50 mM Tris-Cl, pH 7.4, 5 mM NaF, 250 mM NaCl, 5 mM EDTA, 5 mM EGTA, 0.1% (vol/vol) NP40, 5 Mg/ml leupeptin, 10 Mg/ml aprotinin, 10 Mg/ml soybean trypsin inhibitor, 100 MM benzamidine and were spun in an Eppendorf (Hamburg, Germany) centrifuge for 5 min. The supernatant was added to 10 Ml of pl3sucl-beads equilibrated with bead buffer. The tubes were rotated gently for 1 h at 4°. The beads were harvested by centrifugation and washed three times with bead buffer before extraction with 20 Ml of SDS sample buffer and analysis by SDS-PAGE and autQradiography.

Immunoprecipitations and Immunoblotting Mouse anti-p34""2 polyclonal antisera and monoclonal antibodies were raised against the C-terminal two-thirds of Xenopus p34CdC2, starting at an internal Nco I site whose ATG encodes M 85. This clone was kindly provided by John Newport; see Milarski et al. (1991). The Nco I-BamHI fragment was subcloned into pET8c and expressed (as insoluble inclusion bodies) in E. coli BL21(DE3) according to Studier et al. (1990). Monoclonal antibody A17 reacted with p34cdc2 and not with p32cdk2 and was capable of immunoprecipitating cyclins that were associated with p34CdC2 (many we obtained only recognised free p34Cdc2). We typically added 100 Ml of crude culture medium from the myeloma to 5 ,l of translation mix and then recovered the immune complexes with 10-20 Ml of protein A-Sepharose. The procedure for polyclonal antisera was slightly different. The antisera (typically 2 Ml) were first incubated with 20 Ml of protein A-Sepharose suspension (Pharmacia) for 15 min at 40 with gentle mixing. These antibodyloaded beads were washed three times in phosphate-buffered saline, and then mixed with 5 Mil of the [35S]methionine-labeled translation

1282

reactions diluted to a final volume of 400,l with bead buffer. In both cases, the bound proteins were eluted with SDS gel sample buffer and analyzed by SDS-PAGE. Anti-cyclin Al antibodies were the same as described by Kobayashi et al. (1991a). The anti-cdk2 antibody was raised in rabbits against full-length cdk2 expressed in pET3. The antic-myc epitope monoclonal (Evan et al., 1986) was obtained from the ICRF monoclonal production unit. Proteins were transferred electrophoretically onto Immobilon filters (Millipore, Bedford, MA) and PSTAIRE-positive polypeptides were detected with the monoclonal antibody of Yamashita et al., (1992), with the Amersham (Aylesbury, England) ECL detection system.

Histone Hi Kinase Assay To assay the histone Hi kinase activity associated with cyclin A translated in vitro, the CSF extracts were made mRNA-dependent essentially as described by Murray (1991). In brief, the CSF extracts were treated with 0.5 Mg/ml RNase A (Boehringer) for 10 min at 10° followed by one-fortieth volume of 50% (vol/vol) RNasin (Boehringer) for 10 min at 10°. Next, 0.4 mM CaC12, 200 mM sucrose, and 50 yg/ ml calf liver tRNA (Boehringer) were added and the reaction incubated for 15 min at 230. Aliquots were frozen in liquid nitrogen and stored at -80° until use. Aliquots of these nuclease-treated extracts were incubated with 100 added cyclin mRNA and [35S]methionine, and at 30-min inAg/ml tervals, samples were taken for affinity absorption on p13`uc1 beads as described above, except that the buffer in this case was 80 mM Na 3-glycerophosphate, pH 7.3, 20 mM EGTA, 15 mM MgCl2, and 1 mM DTT. The beads were assayed in 10 Al of HI kinase buffer (50 mM Na f-glycerophosphate, pH 7.3, 0.3 mM EDTA, 15 mM MgCl2, 2 mM DTT [Cicirelli et al., 1988]) with 25 AM ATP, 0.25 MCi/Ml y[32PJATP (Amersham PB 10218), and 275 Mg/ml histone Hi (Boehringer). After 30 min at 23°, the reaction was terminated by addition of 25 gl of SDS sample buffer. The samples were analyzed for cyclin synthesis and histone phosphorylation by SDS-PAGE and autoradiography. Activation of Hi kinase by protein A-bovine cyclin A was achieved by addition of 0.4 Ag of wild-type or mutant protein to 25 Ml of interphase Xenopus egg extract. After incubation for 10 min at 21°, 250 ,ul of H 1 kinase buffer was added, followed by 20 Al of IgG-Sepharose slurry (Pharmacia). The samples were rotated at 40 for 1 h, the beads recovered, washed, and assayed for kinase activity as described above.

Xenopus Cultured Cells A line of Xenopus fibroblasts (WAK) were obtained from R.A. Laskey (Cambridge) and propagated in 75% (vol/vol) E4 medium supplemented with 10% (vol/vol) fetal calf serum at 270. At a density of -5 X 10' cells/14-cm plate, the cells were washed with serum-free medium and lysed on ice with buffer containing 50 mM Tris-Cl, 250 mM NaCl, 5 mM EDTA, 50 mM NaF, 0.1% NP40, 10 Mug/ml cytochalasin B, 1 mM PMSF, 1 Mg/ml leupeptin, 2 ,g/iml aprotinin, 10

,g/ml soybean trypsin inhibitor, 15 ,ug/ml benzamidine, and 10 Agl ml pepstatin, pH 7.5. The cells were scraped off and debris removed by centrifugation at 17 000 g for 30 min. The supernatant was used for immunoblotting.

Cyclin Destruction Assays To measure the ability of cyclin A mutants to undergo programmed proteolysis, the test mRNA was translated in "CSF extracts" as described above. Translation was allowed to proceed for 1 h in the presence of [35S]methionine at which point 0.1 mM cycloheximide was added to block further protein synthesis. The sample was divided in two, and one sample was made 0.4 mM in CaCl2, which triggers the destruction of cyclin after a few minutes' lag (Lohka and Maller, 1985; Murray et al., 1989; Shamu and Murray, 1992). Samples were taken for analysis on SDS-PAGE at intervals after adding the CaC12.

Molecular Biology of the Cell

Functional Domains of Cyclin A

The Conservation Plot for Comparing

FXXXVDE sequence do not appear when cyclin A sequences are strictly aligned because of slight differences in the spacing of these elements with respect to the cyclin box and each other. When locally aligned, they give the values shown in Figure 1. Their positions in the plot are those found in frog cyclin Al.

Cyclin A Sequences Figure 1 contains a plot that compares the degree of polypeptide sequence conservation in examples of cyclin A from a variety of species: chicken, clam, cow, frog, fruit fly, hamster, human, limpet, mouse, and sea urchin (Swenson et al., 1986; Lehner and O'Farrell, 1989; Minshull et al., 1990; Wang et al., 1990; van Loon et al., 1991), and see ACKNOWLEDGMENTS). The sequences were aligned by the Intelligenetics Geneworks® (Mountain View, CA) protein alignment module and adjusted by eye. At each position in the sequence, the number of different amino acids was counted. Thus at position 197, all 10 examples of cyclin A have arginine, so the score is 1. At position 204, where the other nine sequences have an L, limpet cyclin A has an M. This position therefore scored 2. These scores were then converted (arbitrarily) as follows: 1 remained 1, 2 became 0.8, 3 became 0.6, 4 became 0.4, and lower scores were transformed into their reciprocal. These numbers were plotted against residue number, and the conservation axis was relabeled as 1, 2 etc. In this way, peaks of conservation and troughs of nonconservation in the sequence comparisons are easily visualised. One additional adjustment was made. The peaks produced by the "destruction box" and the conserved

sty

Eco RV

RESULTS The cDNA for cyclin Al from Xenopus contains unique Sty I, EcoRV, HindIll, and Nsi I sites as shown in Figure 1, which, with Table 1, summarizes most of the deletions and mutations we have made. Figure 1 shows how these mutations map in relation to the sequence conservation shown by a variety of examples of cyclin A. For example, mutant A133 removed all of the N-terminal 133 residues of cyclin Al by fusing the efficient translational start site of influenza virus NS protein to the Hindlll site (Dasso and Jackson, 1989). Other mutants contained internal deletions of varying size and retained some

Hind III

Activity

Nsi

A21 A62 Al 33 Al 44 Al 47 Al 58 Al 61 Al 69 A176 A201 A213 A245 A232

++ ++ ++ ++ ++ ++ ++

CAl4 CA24 CA50 CA79

CA200 Wild type

++

0

Alanine

-p

mutants

AA

RL

1

Y E

A

A

R

D

LK

E

F

PS WY

KY

c

0

(a

0) 0

2

3 4

. . . . . . .I .-u

0

100

200

300

.

400

Residue number

Figure 1. Summary of cyclin A deletions. Full length cyclin Al (wild type) contains 419 amino acids. Solid lines denote the open reading frame of cyclin A, the cyclin box is a black rectangle, and the destruction domain is an open box. The activity of the mutant polypeptides in binding to p34cdc2 (all constructs), activating histone Hi kinase in vitro (wild-type, A21, A161, and 169), or promoting oocyte maturation (see Table 2) is indicated at the right side of each construct. The diagram below represents sequence conservation in the cyclin A family determined and plotted as described in MATERIALS AND METHODS. Peaks indicate conserved residues and troughs lack of conservation at the numbered residues. Vol. 3, November 1992

1283

H. Kobayashi et al.

portion of the correct N-terminus. They are named for their most C-terminal missing residue. The C-terminal deletion series was constructed with the 9E10 epitope from human c-myc replacing the N-terminal 21 amino acids of cyclin Al. All the constructs contained a T7 promoter upstream to allow cell-free synthesis of capped mRNA, and all ended with a termination codon, as indicated in Table 1. The activity of cyclin A mutant mRNAs were tested by microinjection into stage VI Xenopus oocytes, which were incubated and scored for their ability to undergo meiotic maturation. Controls of untreated, progesterone-treated and wild-type cyclin A mRNA-injected oocytes were included in every experiment. The results of this test are given in Table 2. Wild-type cyclin mRNA typically caused maturation in all oocytes in a sample of .20, starting at 2-3 h and completing by 5-6 h. Deletions up to and including A161 were as active as full-length cyclin Al, whereas all deletions from A169 onwards were completely inactive. As shown in Figure 1, the first conserved residue in all known examples of cyclin A is Y 164, closely followed by 1 168. These residues are conserved in the majority of known mitotic cyclin sequences, including cdcl3 of Schizosaccharomyces pombe and the CLB genes of budding yeast. A161 preserves both these residues, whereas they are deleted from A169. We conclude that these residues are necessary for the activity of cyclin A, although we have not tested this point by making point mutations. Deletion Mutants of Cyclin A That Fail to Promote Oocyte Maturation Cannot Bind to p13sucl-Sepharose or to p34cdc2 We next translated several of the cyclin constructs described in Table 1 in the Xenopus cell-free system (Figure 2A), and tested which of them could bind to Sepharose, an affinity resin for p34cdc2 and associated cyclins (Brizuela et al., 1987; Draetta et al., 1989; Labbe et al., 1989; Pondaven et al., 1990). Figure 2B shows that the cyclin A deletion mutants that were active in the oocyte maturation assay bound to p13sucl beads, whereas the inactive mutants failed to bind. It is still not completely clear how p34cdc2 binds to p13SUcl, however, and it was possible that the mutant cyclins might form stable complexes with p34cdc2 that were no longer capable of associating with pl3sucl. Figure 2D shows that this was not the case, because when we used a monoclonal anti-p34cdc2 antibody to immunoprecipitate the various cyclin A constructs, the results exactly followed those of the activity and pl3sucl-binding assays. Constructs A133, A144, and A161 bound to p34Cc,2 whereas A169 and deletions extending beyond the start of the cyclin box did not bind. The C-Terminus of Cyclin A is Required for Binding

pl3ucl-

to p34cdc2 To define the C-terminal end of cyclin A that is necessary for binding to p34CdC2, we constructed a series of

1284

Table 2. Oocyte maturation promoted by N- and C-terminal deletions of cyclin Al Microinjection

Wild-type cyclin Al A21 (myc tagged)

A133 A144

A147 A158

A161 A169

A176 A201 A213 A245

CA14 CA25 CA50

CA200

Total

GVBD

% GVBD

23 24 25 12 13 23 23 21 29 14 10 19 32 26 36 22

22 23 24 12 12 22 23 1 0 0 0 0 2 4 0 0

96 96 96 100 92 96 100 5 0 0 0 0 6 15 0 0

Stage VI oocytes were injected with mRNA corresponding to the constructs shown in Figure 1 and scored for maturation as described in MATERIALS AND METHODS. The ability to promote maturation is lost between residues 161 and 169, and the first 133 residues are wholly dispensable in this assay.

C-terminal deletion mutants (see MATERIALS AND METHODS). The ability of these constructs to bind with companion kinase subunits was tested in the same way as the experiment shown in Figure 2. Immunoprecipitation was performed with the Al 7 monoclonal antip34cdc2 antibody. Figure 3 shows that the shortest Cterminal deletion, lacking only 14 amino acids from its C-terminus, failed to bind to p34cdc2, and removal of 24 and 50 residues did not restore binding. Larger deletions of 79, 139, and 200 residues were similarly inactive. We also made a construct that deleted the 16 C-terminal residues of bovine cyclin A (see Figure 5), which showed no detectable cdc2 binding activity in experiments like the one shown in Figure 3, or histone Hi kinase activity, as Figure 6 shows. When these C-terminal deletions were tested by the oocyte maturation assay (Table 2), we found that CA14 and CA24 had slight residual activity; -10% of the oocytes matured with normal kinetics. This may indicate that the in vivo assay is capable of detecting weaker interactions between p34cdc and cyclin A than the in vitro binding assay. Clearly, however, these C-terminally deleted constructs are seriously impaired in activity.

Activation of Histone Hi Kinase by Mutants of Cyclin A The ability of full-length cyclin Al and the two mutants A161 and A169 were further tested for their ability to activate histone kinase activity in interphase extracts of Molecular Biology of the Cell

Functional Domains of Cyclin A

B

A

Mr

Mr~~~~~~~~~~~ p aN

-

za6lt>;

V-ttEsf;9;N\N9>5 )¢/Kb

p

s

116-

65 cyclin A

-

: B-type

cyclins

40 33-

23

-a

a

-

Bound to p13-Sepharose

Translation

D

C

( 0rANv .cIStA-A

Mr 97 65 58 55

IB-type cyclins

qem

40 33 Sao

Translation

Immunoprecipitation with a-cdc2 antibody

Figure 2. Association of cyclin mutants with p34CdC2. mRNAs corresponding to the indicated cyclin A mutants were translated in Ca2+-treated CSF extract with [35S]methionine. (A) Analysis of the complete reaction by SDS-PAGE and autoradiography. (B) Analysis by adsorption to pl3SUcI Sepharose. B-type cyclins running at -55 kDa (right) provided an internal positive control for recovery. (C) Translation of selected mutants in

a

different extract and (D) Analysis of the binding of these constructs to

Xenopus eggs. Figure 4 shows that there was a gradual increase in histone Hi kinase activity associated with pl3sucl-Sepharose beads prepared from extracts that were incubated with mRNA encoding wild-type cyclin Al or A161, but no increase in kinase activity was observed when A169 or no mRNA was added to the Vol. 3, November 1992

a

monoclonal anti-p34cdc2 antibody Al 7.

reactions. Similar results were obtained when immunoprecipitates of myc-tagged cyclin Al produced by cellfree translation were assayed for their associated histone kinase activity. Wild-type cyclin Al displayed activity (for example, see Figure 9), whereas C-terminal truncations completely lacked it. 1285

H. Kobayashi et al.

B

A

l

41

Activity

C

CPI>Caz G

Frog

C*>

3

PrA-cow A A267 Cow A275 A278 L CA16 Wild Frog type

B-type cyclins

I mmunoprecipitation with a-cdc2 antibody

Translation

Figure 3. C-terminal deletions abolish cyclin A binding to p34cdc2. Cyclin A A21 and C-terminally truncated derivatives were translated in Xenopus egg extract (A) and immunoprecipitated with monoclonal anti-p34cdc2 (B).

Small Deletions in the Cyclin Box Reduce or Abolish Activity of Cyclin A We made another set of internal deletions in Xenopus cyclin Al with its unique Nsi I site or in bovine cyclin A with its unique Acc I site, as indicated in Figure 5. Removal of two or more amino acids at the Nsi I site completely inactivated frog cyclin A, as measured by the oocyte maturation assay (Table 3) or by the in vitro binding assay. The deletion of two amino acids further down the sequence in bovine cyclin A changed the sequence VYITD in the italicized sequence below to VSD, i.e., loss of YI and changing the next T to S: ASKFEEIYIPPEVAEFVYITDDTYTKKQVLRME (underlined letters indicate the residues conserved in cyclin A from clam, fly, frog, and man). This construct (A267) gave 10% residual kinase activity when the construct was added to an interphase extract derived from Xenopus eggs (Figure 6), which we interpret as showing that weak Cyclin A

Al 61

Al 69

no

A

D226A R197A A232 A235 A237 A261 -AA L AVTA

-

A

-

++

+l-

a

Figure 5. The structure of cyclin box mutants in frog and cow cyclin A. The position of the alanine substitutions and internal deletions in the cyclin box are shown to scale. Protein A-cyclin A is missing the first 76 residues of bovine cyclin A and has 264 residues of protein A at the N-terminus that are not shown here. Activity indicated in the right margin comes from binding to p34cdc2 and the ability to activate the histone Hi kinase of p34cdc2.

binding to p34cdc2 gives weak activation. Longer deletions in this location completely abolished the activity of bovine cyclin A. It is noteworthy that so far we have not found

a

mutant cyclin A that can bind to p34cdc2

but not activate it as a protein kinase. It thus appears that, if cyclin A can bind to p34cdc2, it turns on the kinase activity of the complex.

Single Point Mutations of Conserved Residues in the Cyclin Box Abolish the Ability

of Cyclin Al to Bind and Activate p34cdc2 Five residues in the cyclin box are conserved in essentially all known cyclins, including the yeast CLN genes (the major exception is cyclin C). They are (in Xenopus numbering) R197, D226, L241, K252, and E281. So far we have constructed, confirmed by sequencing, and tested R197

--

A and D226

--

A

(Figure 5).

Both of

these point mutations caused complete loss of ability of

mRNA Table 3. Small deletions within the cyclin box abolish oocyte maturation promoted by cyclin Al mRNA Microinjection

0 30 60 90 120 0 30 60 90 120 0 30 60 90 120 0 30 60 90 120 Time (minutes) Figure 4. Histone Hi kinase activity associated with wild type or mutant A161 and A169 cycin A. Mutant or wild-type cyclin A mRNAs were translated in nuclease-treated activated egg extracts, and the histone Hi kinase activity associated with p13sucl beads was measured at various times after the start of the incubation as indicated below each lane.

1286

Progesterone Wild-type cyclin Al A232

A235 Lv237

zv261

Total

GVBD

% GVBD

30 22 24 28 27 25

26 20 0 0 0 0

87 91 0 0 0 0

Stage VI oocytes were injected with mRNA encoding the mutants of Nsi I series whose structures are shown in Figure 5 and Table 1.

Molecular Biology of the Cell

Functional Domains of Cyclin A

the regions in between. Although it was originally found that p32cdk2 did not substitute for cdc2 in S. pombe (Paris et al., 1991), more recent evidence indicates that the human homologue of Egl, now known as cdk2 (cyclin dependent kinase) is capable of such substitution, particularly when certain other mutations were present in the tester strains (Elledge and Spottswood, 1991; Meyerson et al., 1992a).

To test whether Xenopus cyclin Al could associate with p32cdk2 as well as p34cdc2, we co-translated mRNA encoding myc-tagged cyclin Al with cdc2 or cdk2 mRNA in the CSF extract. At the end of the incubation with [35S]methionine, antibodies against cyclin A were added and the immunoprecipitates collected on protein A-Sepharose beads. Figure 7 shows that both p34cdc2 and p32cdk2 were immunoprecipitated with anti-myc or anti-cyclin A antiserum when full length myc-tagged cyclin Al (A21) mRNA or the mRNA for the active A161 construct was present. But when mRNAs encoding the C-terminally truncated cyclin A construct CA14 or the inactive A169 mRNA were used, the binding of the kinase subunits was the same as background reactions in which no cyclin A mRNA was added. Thus mutants that fail to associate with p34cdc2 also cannot bind

Figure 6. Histone Hi kinase activity associated with protein A-cyclin A mutants. Wild-type protein A-cyclin A or the mutant constructs shown in Table 1 and Figure 5 were added to CSF extract that had been activated by addition of Ca2". After 10 min, the cyclin-p34cdc2 complexes were recovered on IgG beads and assayed for histone Hi kinase as described in MATERIALS AND METHODS.

the mutant cyclin Al to form complexes with p34cdc2 in the CSF-extract translation and immunoprecipitation assay.

p32cdk2

Xenopus Cyclin Al Can Bind Both p34cdc2 and p32-cdk Xenopus eggs contain two well-characterized members of the cdc2 family, p34cdc2, and p32cdk2, the latter originally known as Egl, because its mRNA was polyadenylated and translated only during the time of progesterone-induced maturation until the time of fertilization (Paris et al., 1991). The sequences of these two polypeptides are very similar in the regions where all examples of 'cdc2' proteins are conserved and differ in

A

I

We conclude that under the conditions of these assays, when the levels of both cyclin A and p32cdk2 have been increased by translation of added mRNA, cyclin A can bind to either p34cdc2 or to p32cdk2. We should stress that we stand by our previous data (Minshull et al., 1990), which has been confirmed by Fang and Newport (1991) and by Maller and his colleagues (Gabrielli et al., 1992), that in frog egg extracts, no cyclin Al can be detected in association with p32cdk2. Presumably this is partly a matter of relative concentrations

B

Cyclin mRNA

None A21 CAI4 A161 A169

cdc2 + cdk2

+

+

+

+

+ +

None A21 CA14 A161 A169

+

+

Cyclin mRNA +

+

+

+

+

+ +

+

+

+

+

mm

Figure 7. Cyclin A mutants that fail to bind to p34cc cannot bind to p32ck. Cyclin Al mRNA was co-translated with mRNA for either p34CdC2 or p32u'2 in a cell-free translation system composed of equal volumes of Xenopus CSF extract and nuclease-treated rabbit reticulocyte lysate. Reactions were immunoprecipitated with rabbit polyclonal anti-cyclin A antiserum. (A) Complete translation reactions. (B) Immunoprecipitates.

Vol. 3, November 1992

cdc2_--~ _ _

-

Translation

-

\

r

cdk2

Immunoprecipitation with a-cyclin A antibody 1287

H. Kobayashi et al.

because as Figure 8A shows, the levels Of 34Cdc2 are 10-20 times higher than those of the p32cd in normal frog eggs, in agreement with Gabrielli et al., (1992). By contrast, extracts of somatic Xenopus cells (WAK fibroblasts) contain roughly equal concentrations of p34cdc2 and p32cdk2. To get an idea of the relative affinities of p34C c2 and p32c2 subunits for cyclin A, soluble protein A-bovine cyclin A protein produced by expression in bacteria was added to a frog egg extract in which the mRNA for p32cdk2 or p34Cdc' had been translated with added [35S]methionine, and the cyclin (together with any bound proteins) recovered by incubation with IgGSepharose beads. Figure 8B shows that both p34cdc2 and p32cdk2 bound to the beads, but considerably more of the labeled p34cdc2 was retained. Apart from the background of labeled polypeptides produced by translation of endogenous mRNAs, we included negative controls of a truncated cyclin B2 mRNA and a Xenopus PCTAIRE clone (Figure 8, lanes 3 and 4). Compared with p32cdk2 and p34cdc2, these did not bind significantly to the protein A-cyclin A. Because the amount of label in each kinase subunit was adjusted to be the same, this means that the specific activity of the p34cdc2 was roughly onetenth of that of the p32cdk2. Thus even though the cyclin A was added in excess, there was a strong tendency for p34cdc2 to bind in preference to p32cdk2 in the crude egg extract, which may explain why it is normally impossible to find cyclin A associated with p32cdk2 in these extracts.

Cyclin A-p32cdk2 Complexes Possess Histone Hi Kinase Activity To measure the protein kinase activity of p32cdk2 with histone Hi as the substrate, we used a C-terminally cmyc-tagged version (Kobayashi et al., 1991a). The mRNA for this tagged p32cdk2 was translated in the Xenopus egg/reticulocyte lysate mixture either in the presence, or the absence, of wild-type cyclin Al mRNA (Figure 9A, lanes 2 and 3). The C-terminally c-myctagged p32cdk2 construct used for these experiments showed a small increase in mobility when co-translated with cyclin Al mRNA (lane 3), probably associated with its activation (Gu et al., 1992). This shift was not observed when wild-type cdk2 mRNA was used (Figure 7A). When no mRNA was added (the negative control), no kinase activity was associated with the 9E10 anti-cmyc antibody (Figure 9C, lane 1) and only very low activity with polyclonal anti-cdk2 antibody (Figure 9C, lane 3). But as Figure 9C, lane 4 shows, the anti-c-myc antibody carried strong histone Hi kinase activity when p32cdk2_c_myc was translated together with wild-type cyclin Al. When c-myc-cyclin A was translated with no other added mRNA (in which case, the newly translated cyclin A can associate with the endogenous, unlabeled p34cdc2), strong histone Hi kinase activity was present in the 9E10 immunoprecipitate (Figure 9C, lane 1288

A

95% of the endogenous cyclin A mRNA has been ablated with antisense oligonucleotides (Minshull et al., 1991). Assuming that maturation requires active p34cdc2 (an unexamined point, although it would be exceedingly heretical to question it), we argued that the preformed store of cyclin B2 and p34cdc2 must be adequate to perform the whole process. So, how does cyclin A promote maturation? We show here that mutant cyclin A molecules that cannot bind to p34cdc2 fail to promote maturation. This makes it highly probable that cyclin A acts by combining Vol. 3, November 1992

with free p34cdc2 and activating its protein kinase activity. How is it that newly synthesized cyclin A can activate p34cdc2, whereas the maternal stockpile of cyclin B/cdc2 is devoid of activity? We suppose that cyclin A directly activates free p34c c2 in the oocytes, which then tips the balance of a positive feedback loop, leading to activation of the endogenous cyclin B2-p34cdc2. Precisely how this occurs is not yet clear. As recently discussed in some detail by Devault et al., (1992), it is presumably necessary either to activate the tyrosine 15 phosphatase encoded by cdc25, or to inhibit the tyrosine 15 kinase encoded by a weel-like enzyme, or both, to activate preformed cyclin B-p34cdc2 protein kinase (Dunphy and Newport, 1989; Dunphy and Kumagai, 1991; Gautier et al., 1991). Further work is necessary to discover what happens in frog oocytes.

Why Did Previous Studies Fail to Detect Cyclin A-

p32cdk2 Association in Xenopus Extracts? We were at first surprised to find that Xenopus p32cdk2 could form complexes with cyclin A, considering that studies in three other laboratories besides our own failed to detect newly synthesized cyclin A bound to p32cdk2 (Minshull et al., 1990; Fang and Newport, 1991; Paris et al., 1991; Gabrielli et al., 1992). More recently, data from mammalian cells has shown that cyclin A is often found predominantly associated with cdk2 (Tsai et al., 1991; Desai et al., 1992). The data in this paper clearly show that Xenopus cyclin A and p32cdk2 can, like their mammalian counterparts, bind tightly together and produce active histone Hi kinase. However, there are potentially important differences in the way our more recent experiments were performed and the situation in intact Xenopus eggs and embryos. First, to obtain high level translation of cdk2 mRNA, we found it necessary to add reticulocyte lysate to the Xenopus extract; it is thus possible that the presence of reticulocyte lysate may promote the interaction in some way, although it does not do so in the case of inactive mutants, like CA14. More relevant, perhaps, is that the concentration of p32cdk2 is normally much less (perhaps -5-10%) of p34cdc2 in frog extract (see Figure 8), and the concentration of cyclin A is probably lower still. In the cellfree translation experiments, the concentrations of both components are much increased, which tends to promote interaction. There is very little cyclin A in normal eggs, and they contain a large excess of p34cdc2, So it is very likely that most of the cyclin is 'mopped up' by the p34cdc2. The experiment shown in Figure 8 suggests that in the frog egg extract, the relative affinities of p34cdc2 for cyclin A is higher than that of p32cdk2 for cyclin A. These results by no means exclude the possibility that p32cdk2 has other, as yet unidentified, cyclin or cyclin-like subunits. What is perhaps more puzzling is that in human fibroblasts, cyclin A is mainly found associated with p32cdk2 and not with p34cdc2 (Tsai et al., 1291

H. Kobayashi et al.

1991). One possible explanation is that the cyclin A found in eggs is different from that found in somatic cells, and we have preliminary evidence that this is true. It will be necessary to design experiments to compare the relative affinities of cyclins Al (the oocyte cyclin A) and A2 (the presumptive somatic cyclin A) for various kinase subunits. Do Some Mutations Cause Misfolding? Both the reviewers of the submitted version of this paper raised the following question: how do we know that proteins with greatly altered sequences fold up correctly? Could nonbinding of cyclin A to p34cdc2 be caused by gross misfolding, rather than deletion of, or alterations in key interacting residues in the subunit interface domain? We did not discuss this question because we thought it was very difficult, if not impossible, to answer in the absence of structural assays that are independent of functional ones. In the case of point mutations in the cyclin box, we would argue that single alanine replacements are generally well tolerated (see Gibbs and Zoller, 1991), but deletions of two amino acids might, we imagine, cause large changes in conformation. At the same time, large deletions of up to 133 amino acids in the N-terminal 161 residues of cyclin A are perfectly tolerated and presumably do not induce misfolding, because they are able to bind and to activate p34cdc and p32cdk2. We are more puzzled by the failure of the short C-terminal truncations of cyclin A to bind to p34cdc2, given the variations in sequence between different members of the cyclin family in this region of the molecule. We are exploring the use of mild protease digestion, which may provide a way to detect gross structural alterations. In truth, however, solving the crystal structure of cyclin A-p34cdc2 kinase is the only sure way to obtain precise, accurate and detailed information.

ACKNOWLEDGMENTS Mary Dasso constructed the first cyclin A mutant, A133, from which many subsequent constructs were derived, and Roy Golsteyn made the first batches, of p133uc' from a clone supplied by Paul Nurse. Dr. Takeharu Nishimoto has given generous support and encouragement, and Mark Carrington, Erich Nigg, Katsumi Yamashita, and Matt Winkler provided us with the sequences of mouse, chicken, hamster, and sea urchin cyclin A before publication. We are very grateful to the staff of the Animal Unit, Cell Production facilities, and Oligonucleotide synthesis unit at ICRF Clare Hall for their expert work in producing the antibodies, cell extracts, and oligonucleotides used here. HK was supported by a Fellowship from the JSPS; J.P.A. by a scholarship from the Stiftung Stipendienfond des VCI e.V., and at the start, the work was funded by the CRC and the Wellcome Trust.

REFERENCES Brizuela, L., Draetta, G., and Beach, D. (1987). p135uc1 acts in the fission yeast cell division cycle as a component of the p34cdc2 protein kinase. EMBO J. 6, 3507-3514. Cicirelli, M.F., Pelech, S.L., and Krebs, E.G. (1988). Activation of multiple protein kinases during the burst in protein phosphorylation that precedes the first meiotic cell division in Xenopus oocytes. J. Biol. Chem. 263, 2009-2019.

1292

Colman, A. (1984). Translation of eukaryotic messenger RNA in Xenopus oocytes. In: Transcription and Translation, ed. B.D. Hames and S.J. Higgins, Oxford, U.K.: IRL Press, 271-302. Dasso, M.C., and Jackson, R.J. (1989). Efficient initiation of mammalian mRNA translation at a CUG codon. Nucleic Acids Res. 17, 64856497. Desai, D., Gu, Y., and Morgan, D.O. (1992). Activation of human cyclin-dependent kinases in vitro. Mol. Biol. Cell 3, 571-582. Devault, A., Fesquet, D., Cavadore, J.-C., Garrigues, A.-M., Labbe, J.-C., Lorca, T., Picard, A., Philippe, M., and Doree, M. (1992). Cyclin A potentiates maturation-promoting factor activation in the early Xenopus embryo via inhibition of the tyrosine kinase that phosphorylates CDC2. J. Cell Biol. 118, 1109-1120. Draetta, G., Luca, F., Westendorf, J., Brizuela, L., Ruderman, J., and Beach, D. (1989). cdc2 protein kinase is complexed with both cyclin A and B: evidence for proteolytic inactivation of MPF. Cell 56, 829838. Ducommun, B., Brambilla, P., F6lix, M.A., Franza, B.R.J., Karsenti, E., and Draetta, G. (1991). cdc2 phosphorylation is required for its interaction with cyclin. EMBO J. 10, 3311-3319. Dunphy, W., and Newport, J. (1989). Fission yeast p13 blocks mitotic activation and tyrosine dephosphorylation of the Xenopus cdc2 protein kinase. Cell 58, 181-191. Dunphy, W.G., Brizuela, L., Beach, D., and Newport, J. (1988). The Xenopus cdc2 protein is a component of MPF, a cytoplasmic regulator of mitosis. Cell 54, 423-431. Dunphy, W.G., and Kumagai, A. (1991). The cdc25 protein contains an intrinsic phosphatase activity. Cell 67, 189-196. Elledge, S.J., and Spottswood, M.R. (1991). A new human p34 protein kinase, CDK2, identified by complementation of a cdc28 mutation in Saccharomyces cerevisiae, is a homolog of Xenopus Egl. EMBO J. 10, 2653-2660. Evan, G.I., Hancock, D.C., Littlewood, T., and Pauza, C.D. (1986). Characterization of the human c-myc protein using antibodies prepared against synthetic peptides. Ciba Found. Symp. 119, 245-263. Fang, F., and Newport, J.W. (1991). Evidence that the G1-S and G2M transitions are controlled by different cdc2 proteins in higher eukaryotes. Cell 66, 731-742. Featherstone, C., and Russell, P. (1991). Fission yeast pl0 7eIl mitotic inhibitor is a tyrosine/serine kinase. Nature 349, 808-811. Fleig, U.N., and Gould, K. (1991). Regulation of cdc2 activity in Schizosaccharomyces pombe. Semin. Cell Biol. 2, 195-204. Gabrielli, B.G., Roy, L.M., Gautier, J., Philippe, M., and Maller, J.L. (1992). A cdc2-related kinase oscillates in the cell cycle independently of cyclins G2/M and cdc2. J. Biol. Chem. 267, 1969-1975. Gautier, J., Solomon, M.J., Booher, R.N., Bazan, J.F., and Kirschner, M.W. (1991). cdc25 is a specific tyrosine phosphatase that directly activates p34cdc2. Cell 67, 197-211. Gibbs, C.S., and Zoller, M.J. (1991). Rational scanning mutagenesis of a protein kinase identifies functional regions involved in catalysis and substrate interactions. J. Biol. Chem. 266, 8923-8931. Girard, F., Strausfeld, U., Fernandez, A., and Lamb, N.J. (1991). Cyclin A is required for the onset of DNA replication in mammalian fibroblasts. Cell 67, 1169-1179. Glotzer, M., Murray, A.W., and Kirschner, M.W. (1991). Cyclin is degraded by the ubiquitin pathway. Nature 349, 132-138. Gould, K.L., Moreno, S., Owen, D.J., Sazer, S., and Nurse, P. (1991). Phosphorylation at Thrl67 is required for Schizosaccharomyces pombe p34cdC2 function. EMBO J. 10, 3297-3309. Molecular Biology of the Cell

Functional Domains of Cyclin A

Gould, K.L., and Nurse, P. (1989). Tyrosine phosphorylation of the fission yeast cdc2+ protein kinase regulates entry into mitosis. Nature 342, 39-45. Gu, Y., Rosenblatt, J., and Morgan, D.O. (1992). Cell cycle regulation of CDK2 activity by phosphorylation of threonine 160 and tyrosine 15. EMBO J. 11, 3885-4005. Hadwiger, J.A., Wittenberg, C., Richardson, H.E., de Barros-Lopes, M., and Reed, S.I. (1989). A novel family of cyclin homologs that control Gl in yeast. Proc. Natl. Acad. Sci. USA 86, 6255-6259. Horton, R.M., and Pease, L.R. (1991). Recombination and mutagenesis of DNA sequences using PCR. In: Directed Mutagenesis, ed. M.J. McPherson, Oxford: IRL Press at OUP, 217-247. Kobayashi, H., Golsteyn, R., Poon, R., Stewart, E., Gannon, J., Minshull, J., Smith, R., and Hunt, T. (1991a). Cyclins and their partners during Xenopus oocyte maturation. Cold Spring Harbor Symp. Quant. Biol. 56, 437-447. Kobayashi, H., Minshull, J., Ford, C., Golsteyn, R., Poon, R., and Hunt, T. (1991b). On the synthesis and destruction of A- and B-type cyclins during oogenesis and meiotic maturation in Xenopus laevis. J. Cell Biol. 114, 755-765. Kumagai, A., and Dunphy, W.G. (1992). Regulation of the cdc25 protein during the cell cycle in Xenopus extracts. Cell 70, 139-151. Labbe, J.-C., Capony, J.-P., Caput, D., Cavadore, J.-C., Derancourt, J., Kaghdad, M., Lelias, J.-M., Picard, A., and Doree, M. (1989). MPF from starfish oocytes at first meiotic metaphase is a heterodimer containing one molecule of cdc2 and one molecule of cyclin B. EMBO J. 8, 3053-3058. Lehner, C.F., and O'Farrell, P.H. (1989). Expression and function of Drosophia cyclin A during embryonic cell cycle progression. Cell 56, 957-968. Lehner, C.F., and O'Farrell, P.H. (1990). Drosophila cdc2 homologs: a functional homolog is coexpressed with a cognate varient. EMBO J. 9, 3573-3581. Leopold, P., and O'Farrell, P.H. (1991). An evolutionarily conserved cyclin homolog from Drosophila rescues yeast deficient in Gl cyclins. Cell 66, 1207-1216. Lew, D.J., Dulic, V., and Reed, S.I. (1991). Isolation of three novel human cyclins by rescue of Gl cyclin (Cln) function in yeast. Cell 66, 1197-1206. Lohka, M., and Maller, J. (1985). Induction of nuclear envelope breakdown chromosome condensation, and spindle formation in cellfree extracts. J. Cell Biol. 101, 518-523. Lorca, T., Devault, A., Colas, P., Van Loon, A., Fesquet, D., Lazaro, J.B., and Doree, M. (1992). Cyclin A-Cys4l does not undergo cell cycle-dependent degradation in Xenopus extracts. FEBS Letts. 306, 9093. Luca, F.C., and Ruderman, J.V. (1989). Control of programmed cyclin destruction in a cell-free system. J. Cell Biol. 109, 1895-1909. Luca, F.C., Shibuya, E.K., Dohrmann, C.E., and Ruderman, J.V. (1991). Both cyclin A delta 60 and B delta 97 are stable and arrest cells in Mphase, but only cyclin B delta 97 turns on cyclin destruction. EMBO J, 10, 4311-4320. Maller, J.L. (1991). Mitotic control. Curr. Opin. Cell. Biol. 3, 269-275. Matsushime, H., Roussel, M.F., Ashmun, R.A., and Sherr, C.J. (1991). Colony-stimulating Factor 1 regulates novel cyclins during the Gl phase of the cell cycle. Cell 65, 701-713. Meyerson, M., Enders, G.H., Wu, C.L., Su, L.K., Gorka, C., Nelson, C., Harlow, E., and Tsai, L.H. (1992). A family of human cdc2-related

protein-kinases. EMBO J. 11, 2909-2917. Vol. 3, November 1992

Milarski, K.L., Dunphy, W.G., Russell, P., Gould, S.J., and Newport, J.W. (1991). Cloning and characterization of Xenopus cdc2. Cold Spring Harbor Symp. Quant. Biol. 56, 377-384. Millar, J.B., and Russell, P. (1992). The cdc25 M-phase inducer: an unconventional protein phosphatase. Cell 68, 407-410. Minshull, J., Golsteyn, R., Hill, C., and Hunt, T. (1990). The A- and B-type cyclin-associated cdc2-kinases in Xenopus turn on and off at different times in the cell cycle. EMBO J. 9, 2865-2875. Minshull, J., Murray, A., Colman, A., and Hunt, T. (1991). Xenopus oocyte maturation does not require new cyclin synthesis. J. Cell Biol. 114, 767-772. Morla, A., Draetta, G., Beach, D., and Wang, J.Y.J. (1989). Reversible tyrosine phosphorylation of cdc2: dephosphorylation accompanies activation during entrance into mitosis. Cell 58, 193-203. Motokura, T., Bloom, T., Kim, H.G., Jiippner, H., Ruderman, J.V., Kronenberg, H.M., and Arnold, A. (1991). A novel cyclin encoded by a bcll-linked candidate oncogene. Nature 350, 512-515. Murray, A.W. (1991). Cell cycle extracts. In: Xenopus laevis: Practical Uses in Cell and Molecular Biology, ed. B.K. Kay and H.B. Peng, San Diego: Academic Press, 573-597. Murray, A.W., Solomon, M.J., and Kirschner, M.W. (1989). The role of cyclin synthesis and degradation in the control of maturation promoting factor activity. Nature 339, 280-286. Nielsen, D.A., and Shapiro, D.J. (1986). Preparation of capped RNA transcripts using T7 RNA polymerase. Nucl. Acids Res. 14, 5936. Nugent, J.H., Alfa, C.E., Young, T., and Hyams, J.S. (1991). Conserved structural motifs in cyclins identified by sequence analysis. J. Cell Sci. 99, 669-674. Nurse, P. (1990). Universal control mechanism regulating cell cycle timing of M-phase. Nature 344, 503-508. Pagano, M., Pepperkok, R., Verde, F., Ansorge, W., and Draetta, G. (1992). Cyclin A is required at 2 points in the human cell-cycle. EMBO J. 12, 961-971. Paris, J., Le Guellec, R., Couturier, A., Le Guellec, K., Omilli, F., Camonis, J., MacNeill, S., and Philippe, M. (1991). Cloning by differential screening of a Xenopus cDNA coding for a protein highly homologous to cdc2. Proc. Natl. Acad. Sci. USA 88, 1039-1043. Parker, L.L., Atherton-Fessler, S., Lee, M.S., Ogg, S., Falk, J., Swenson, K.I., and Piwnica-Worms, H. (1991). Cyclin promotes the tyrosine phosphorylation of p34cdc2 in a weel+ dependent manner. EMBO J. 10, 1255-1263. Parker, L.L., Atherton-Fessler, S., and Piwnica-Worms, H. (1992). p 107w"' is a dual-specificity kinase that phosphorylates p34cdc2 on tyrosine-15. Proc. Natl. Acad. Sci. USA 89, 2917-2921. Pines, J., and Hunter, T. (1990). p34cdc2: the S and M kinase? new biol. 2, 389-401. Pondaven, P., Meijer, L., and Beach, D. (1990). One-step affinity purification of starfish cdc2/cyclin complex: in vitro activation of the kinase with phosphatase. Genes Dev. 4, 9-17. Roy, L.M., Swenson, K.I., Walker, D.H., Gabrielli, B.G., Li, R.S., Piwnica, W.H., and Maller, J.L. (1991). Activation of p34cdc2 kinase by cyclin A. J. Cell Biol. 113, 507-514. Shamu, C.E., and Murray, A.W. (1992). Sister chromatid separation in frog egg extracts requires DNA topoisomerase II activity during anaphase. J. Cell. Biol. 117, 921-934. Solomon, M.J., Glotzer, M., Lee, T.H., Philippe, M., and Kirschner, M.W. (1990). Cyclin activation of p34cdc2. Cell 63, 1013-1024. Strausfeld, U., Labbe, J.C., Fesquet, D., Cavadore, J.C., Picard, A., Sadhu, K., Russell, P., and Doree, M. (1991). Dephosphorylation and

1293

H. Kobayashi et al.

activation of a p34cdc2/cyclin B complex in vitro by human CDC25 protein. Nature 351, 242-245. Studier, F.W., Rosenberg, A.H., Dunn, J.J., and Dubendorff, J.W. (1990). Use of T7 RNA polymerase to direct expression of cloned genes. In: Methods in Enzymology, ed. D.V. Goeddel, San Diego: Academic Press, vol. 185, pp. 60-89. Swenson, K.I., Farrell, K.M., and Ruderman, J.V. (1986). The clam embryo protein cyclin A induces entry into M phase and the resumption of meiosis in Xenopus oocytes. Cell 47, 861-870. Tsai, L.-H., Harlow, E., and Meyerson, M. (1991). Isolation of the human cdk2 gene that encodes the cyclin A- and adenovirus E1Aassociated p33 kinase. Nature 353, 174-177.

1294

van Loon, A.E., Colas, P., Goedemans, H.J., Neant, I., Dalbon, P., and Guerrier, P. (1991). The role of cyclins in the maturation of Patella vulgata oocytes. EMBO J. 10, 3343-3349. Wang, J., Chenivesse, X., Henglein, B., and Br&chot, C. (1990). Hepatitis B virus integration in a cyclin A gene in a hepatocellular carcinoma. Nature 343, 555-557. Xiong, Y., Connolly, T., Futcher, B., and Beach, D. (1991). Human Dtype cyclin. Cell 65, 691-699. Yamashita, M., Fukada, S., Yoshikuni, M., Bulet, P., Hirai, T., Yamaguchi, A., Lou, Y.H., Zhao, Z., and Nagahama, Y. (1992). Purification and characterization of maturation-promoting factor in fish. Dev. Biol. 149, 8-15.

Molecular Biology of the Cell